MINING WEB SEARCH USER BEHAVIOR TO ENHANCE WEB SEARCH RELEVANCE

摘要

The system and method for user preference are estimated by the automatic interpretation of user behavior. User behavior component associated with search engine can be with the collective behavior of automatic interpretation user (for example, web search user). Such feedback component may include user behavior characteristics and prediction model (for example). Robust to noise from user behavior component, can reside in being interacted with the user of search result observed (for example, malice and/or unreasonable User Activity).